arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.17819cs.LGcs.AI
收藏

自动自适应共形风险控制

Automatically Adaptive Conformal Risk Control

Paris-Saclay University(巴黎萨克雷大学);CNRS(法国国家科学研究中心);University of California, Berkeley(加州大学伯克利分校);Laboratoire Interdisciplinaire des Sciences du Numerique(数字科学跨学科实验室);Capgemini Invent France(凯捷创新法国公司);ENSIIE(法国国立统计与信息工程学院);INRIA Paris(法国国家信息与自动化研究所巴黎分部)

Vincent Blot, Anastasios N Angelopoulos, Michael I Jordan, Nicolas J-B Brunel

AI总结:

提出一种自动自适应共形风险控制框架,通过数据驱动确定条件化函数类,实现对统计风险的近似条件控制,在回归和分割任务中比传统方法精度更高。

摘要:

科学技术日益需要有效的机制来确保黑盒机器学习算法提供可靠且受控的性能。这些性能保证理想情况下应基于输入条件成立——即无论输入是什么,性能保证至少应近似成立。然而,除了种族和性别等程式化的离散分组之外,合适的条件化概念可能难以定义。例如,在图像分割等问题中,我们希望不确定性反映测试样本的内在难度,但这可能难以通过条件事件来捕捉。基于Gibbs等人[2023]的近期工作,我们提出了一种方法,通过适应测试样本的难度,实现对统计风险(损失函数的期望值)的近似条件控制。我们的框架超越了传统的基于用户提供的条件事件的条风险控制,转向算法化、数据驱动地确定合适的条件化函数类。我们将此框架应用于各种回归和分割任务,实现了对模型性能的更细粒度控制,并证明通过持续监测和调整这些参数,与传统风险控制方法相比,我们可以实现更高的精度。

英文摘要:

Science and technology have a growing need for effective mechanisms that ensure reliable, controlled performance from black-box machine learning algorithms. These performance guarantees should ideally hold conditionally on the input-that is the performance guarantees should hold, at least approximately, no matter what the input. However, beyond stylized discrete groupings such as ethnicity and gender, the right notion of conditioning can be difficult to define. For example, in problems such as image segmentation, we want the uncertainty to reflect the intrinsic difficulty of the test sample, but this may be difficult to capture via a conditioning event. Building on the recent work of Gibbs et al. [2023], we propose a methodology for achieving approximate conditional control of statistical risks-the expected value of loss functions-by adapting to the difficulty of test samples. Our framework goes beyond traditional conditional risk control based on user-provided conditioning events to the algorithmic, data-driven determination of appropriate function classes for conditioning. We apply this framework to various regression and segmentation tasks, enabling finer-grained control over model performance and demonstrating that by continuously monitoring and adjusting these parameters, we can achieve superior precision compared to conventional risk-control methods.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑